Hi Lanna -
It's been a few weeks since we've chatted. I hope all is going well for you!
WRT color calibration:
a) I have never heard any good arguments for setting the target colors of the curves/levels eyedroppers to anything except their default settings (ie, changing their "calibration"). My philosophy is that if you have to do this to get a good looking print, it's because you haven't set up your color management procedures correctly. For example, if you find you get better looking highlights in your printed output if you target pure white to 248,248,248, it probably means that either your monitor or printer profile is in error, and all you are doing is an ad hoc work-around.
b) Unless you include a MacBeth color reference chart (or equivalent) in each image that you shoot, it's almost pointless to work by the numbers, at least in this way. And, even if you do include a color reference chart there is a much better method (which I can describe later) to get scientifically accurate colors than trying to do so by "re-calibrating" the eyedroppers.
c) If you don't have a set of color reference squares in each frame, buy a good monitor (ie, NOT a laptop), and a good hardware calibration system, and adjust the colors by eye. If you are not confident about your ability to judge how colors should look, search the web and save a folder full of images of similar subject matter that you think have good color. Then, when you are editing your own images, constantly go back and forth between your image and the good example images. Audio engineers and producers to exactly the same thing when they are mixing down a song -- they go back and forth between the track they are working on and a set of high quality reference audio tracks, and play them through exactly the same system, same (audio) monitors, etc.
d) You are certainly correct to be concerned about the possibility of out-of-gamut colors. Note: in several places, you referred to these as "out of gamma". I hope this was just a typo. There is a profound difference between the meanings and implications of "gamma" and "gamut".
Predominently, out-of-gamut problems occur at the printing step, much less frequently as a monitor / screen display problem.
The correct way to handle out-of-gamut colors is to obtain a color profile from your printer (ie, for one particular combination of machine, inkset, paper, etc.), install it on your system, and turn on soft proofing and out-of-gamut warnings. You don't get rid of this problem by making changes in the curves or levels eyedroppers.
Also, out-of-gamut problems typically occur with extremely saturated colors. If I remember correctly from previous threads, most of your pix were of weakly blue-gray fish. These are *very* unlikely to be out-of-gamut for any reasonable printer. What makes you think you have this problem?
e) You used the phrase, "neutralizing colors". This phrase means changing the color balance so that an area of your image that was colored, now is a neutral gray. My first question to you would be: How do you know that a particular area of your image is truly a neutral gray? Should the skin of a shark really be brought to perfectly neutral gray, or is it more of a steel-gray?
My second comment about "neutralizing" is that completely "neutralized" colors typically don't look natural or attractive, and almost always throw off other areas of the image. As a little example of this, take two pix of a sunset, one with the automatic white balance in your camera off, and a second shot with it on. The AWB is very similar to "neutralizing" color in post processing. I'm quite sure you will find that you will prefer the image shot without the AWB. Extending this concept to your underwater pix, you should consider very carefully whether viewers might actually prefer a bit of blue-green to remain in your underwater photos. They could easily look too sterile with a theoretically perfect white balance.
f) WRT selecting RGB vs CMYK vs LAB, I'm completely in agreement with Larry (ALB) on this. The big advantage of LAB color is that changes in the luminosity don't change the colors. However, it can be very tricky and non-intuitive to make color changes in LAB. Instead of using LAB, I would suggest you go back to RGB, and become familiar with the luminosity blending mode (in layer properties). Just like LAB, this allows you to change luminosity values without changing colors, but without any of the complexities and difficulties of working in LAB.
The next logical question has to be, which of the various RGB spaces should you use, eg, sRGB, Adobe RGB, ProPhoto RGB, etc. Given the questions you are asking, to be honest, it sounds like you are just getting started in the area of color management, so my very strong suggestion would be to work in sRGB, a nice, conservative, narrow gamut space that will tend to protect you from making wild errors in colors. Once you master sRGB (which can take years ;-) ), if you have some fishy subjects that are intensely saturated and are clearly out of gamut, then think of moving to a wider gamut space such as Adobe RGB.
There is a lot to digest in the above material, but feel free to come back with any questions you might have.
Best regards,
Tom
